[Caml-list] Re: Syntax Changes in OCaml

[Caml-list] Re: Syntax Changes in OCaml

[dbmcclain]

I had the distinct impression that the latest release of OCaml already
imposes a (minor) syntax change in that tags on library function arguments
are now mandatory. Is this not the case?

My (mis?)understanding has already prevented my adoption of the latest
version simply because of the effort required to port 20 KLOC of OCaml to
this new imposition. If this is not the case, then I will gladly incorporate
the newest version in my system.

But this greater discussion has me wondering about whether you are talking
of more egregious changes to the language syntax.... If so, it will
certainly push me as a user to consider other alternatives. I don't wish to
spend my time rewriting old code to keep up with a moving target language...

- David McClain, Sr. Scientist, Raytheon Missile Systems Co., Tucson, AZ
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Markus Mottl]

On Mon, 04 Feb 2002, dbmcclain wrote:
> But this greater discussion has me wondering about whether you are
> talking of more egregious changes to the language syntax.... If so,
> it will certainly push me as a user to consider other alternatives. I
> don't wish to spend my time rewriting old code to keep up with a moving
> target language...

There may be a misunderstanding here: the problem with the recent changes,
even if they were rather innocuous, is that you really have to adapt your
sources. At least for me this was not a big deal, because I had mostly
avoided labels so far, but it would be a nightmare if large parts of
the syntax changed.

However, a shift to revised (or another) syntax as default wouldn't be
all this bad: the preprocessor would still handle "normal" syntax, which
means that you'd only have to feed your existing sources through the
preprocessor. This can be done by changing only one line in your Makefile.

You could even continue writing and maintaining code in "normal" syntax,
similar to those few people who are currently using revised syntax as
alternative - without problems other than "social" ones.

The only problem I can imagine for you is that you have to pass
maintainance of your source base to a young programmer who has only been
trained on revised syntax. But even then I don't think that this will be
too serious, because fixing bugs in a slightly different syntax isn't all
this difficult.  Mixing revised and normal syntax (in different files, of
course) is absolutely not a problem either: just add a comment to the top
of your files so that the right preprocessor can be chosen automagically
(as is being supported by OcamlMakefile, for example).

A sudden transition to revised syntax would probably cost me less than
a week, most of it due to getting used to the new syntax. I don't know
for sure, but this may be true for the majority of users.

Regards,
Markus Mottl

-- 
Markus Mottl                                             markus@oefai.at
Austrian Research Institute
for Artificial Intelligence                  http://www.oefai.at/~markus
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Daniel de Rauglaudre]

Hi,

On Tue, Feb 05, 2002 at 12:23:35PM +0100, Markus Mottl wrote:

> However, a shift to revised (or another) syntax as default wouldn't be
> all this bad: the preprocessor would still handle "normal" syntax, which
> means that you'd only have to feed your existing sources through the
> preprocessor. This can be done by changing only one line in your Makefile.

There is another point preventing us to change the syntax: there is no
consensus here about how the syntax must be. I made my choices with
the revised syntax, but somebody says that "this construction is ugly",
somebody else says that "that one (another one) is weird", and so on...
No convergence. And people ends with "pfff... this is *only* syntax".

If the architects don't agree of how the house must be, there is no
chance that the house be built. Therefore there is no plan to propose
or impose (like you would like) a new syntax.

But there is no problem of syntax: if you consider syntax as important,
and don't want to have problems with it, you can use the revised syntax.
If you consider that object programming is very important, use objects.
If you can't live without labels, use labels.

I don't want that people imposes me to use objects and labels, and I
shall not impose people to use the revised syntax.

The language is powerful, it is the reason why we can propose many
features and more and more. Yes it is a problem: it is like the
different window managers under X window, against the standard of
MsWindows of Apple. What is the better way? I don't know.

Here we have tried to improve the system, and this is therefore
something "like X windows", with the drawback that there are several
window-or-syntax managers.

Question: what is the better window manager under X window?

-- 
Daniel de RAUGLAUDRE
daniel.de_rauglaudre@inria.fr
http://cristal.inria.fr/~ddr/
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Jérôme Marant]

On Tue, Feb 05, 2002 at 12:23:35PM +0100, Markus Mottl wrote:
> On Mon, 04 Feb 2002, dbmcclain wrote:
> > But this greater discussion has me wondering about whether you are
> > talking of more egregious changes to the language syntax.... If so,
> > it will certainly push me as a user to consider other alternatives. I
> > don't wish to spend my time rewriting old code to keep up with a moving
> > target language...
> 
> There may be a misunderstanding here: the problem with the recent changes,
> even if they were rather innocuous, is that you really have to adapt your
> sources. At least for me this was not a big deal, because I had mostly
> avoided labels so far, but it would be a nightmare if large parts of
> the syntax changed.

  Isn't the -nolabels option here for backward compatibility?

  Cheers,

-- 
Jérôme Marant
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Pixel]

Daniel de Rauglaudre <daniel.de_rauglaudre@inria.fr> writes:

> The language is powerful, it is the reason why we can propose many
> features and more and more. Yes it is a problem: it is like the
> different window managers under X window, against the standard of
> MsWindows of Apple. What is the better way? I don't know.
> 
> Here we have tried to improve the system, and this is therefore
> something "like X windows", with the drawback that there are several
> window-or-syntax managers.
> 
> Question: what is the better window manager under X window?

i don't think the comparison is valid, at least if you want to share the code
(whereas you don't need to share your window manager)
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Jacques Garrigue]

From: "dbmcclain" <dbmcclain@email.msn.com>

> I had the distinct impression that the latest release of OCaml already
> imposes a (minor) syntax change in that tags on library function arguments
> are now mandatory. Is this not the case?

It is not a syntax change (the grammar didn't change), but a
typechecker change, which only concerns people who use labels.
(We hoped they were flexible enough to accept this integrating change.)
Labels are required in some cases, but there are no labels in the
standard library.

> My (mis?)understanding has already prevented my adoption of the latest
> version simply because of the effort required to port 20 KLOC of OCaml to
> this new imposition. If this is not the case, then I will gladly incorporate
> the newest version in my system.

Indeed this is a misunderstanding: ocaml 3.04 is fully compatible with
ocaml 2.04. Non-labeled programs do not require any change.
(Actually, there is one change in the grammar for stream parsers, and
you have to use camlp4 for them, but this is really all.)

Keep with us,

Jacques Garrigue
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Markus Mottl]

On Tue, 05 Feb 2002, Daniel de Rauglaudre wrote:
> There is another point preventing us to change the syntax: there is
> no consensus here about how the syntax must be. I made my choices with
> the revised syntax, but somebody says that "this construction is ugly",
> somebody else says that "that one (another one) is weird", and so on...
> No convergence. And people ends with "pfff... this is *only* syntax".

When I used the term "revised syntax" I was implicitly including possible
changes to satisfy the majority. And I'd never say "this is *only*
syntax". "It's *only* syntax" that allows you to actually write down
your programs at all.

> If the architects don't agree of how the house must be, there is no
> chance that the house be built. Therefore there is no plan to propose
> or impose (like you would like) a new syntax.

I'd suggest that the "revisionists" should publically discuss how
they would want syntax to evolve. Other people will surely follow the
discussion and comment on it.

> I don't want that people imposes me to use objects and labels, and I
> shall not impose people to use the revised syntax.

To some extent this is comparable, I agree: if one isn't used to
e.g. objects (or higher-order modules, for example), understanding
other people's code may become even more difficult than if it were in
another syntax.

But there is also a difference: if I decided to use objects from
today on, no automatic tool will get reasonably readable code in, say,
"module style" out of my sources, which is not true for purely syntactic
differences.

Therefore, I'd propose that pretty-printing technology be even further
improved, especially what concerns adaptability to user preferences. Then
it would become really easy for people to switch to the syntax they like
rather than having to stick to the one they started out with or the one
which is considered "standard".

Regards,
Markus Mottl

-- 
Markus Mottl                                             markus@oefai.at
Austrian Research Institute
for Artificial Intelligence                  http://www.oefai.at/~markus
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Doug Bagley]

Daniel de Rauglaudre wrote:
> Question: what is the better window manager under X window?

One that allows you to customize its behavior with OCaml, as Sawfish
allows you to customize with Lisp. That sounds like a fun project to
me! :)

cheers,
doug
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr

Re: [Caml-list] Re: Syntax Changes in OCaml

[Remi VANICAT]

doug@bagley.org (Doug Bagley) writes:

> Daniel de Rauglaudre wrote:
> > Question: what is the better window manager under X window?
> 
> One that allows you to customize its behavior with OCaml, as Sawfish
> allows you to customize with Lisp. That sounds like a fun project to
> me! :)

you should look to the Efun/GwML package
(http://pauillac.inria.fr/para/cdrom/prog/unix/efuns/eng.htm). 
GwML is a windows manager written in ocaml, an that is able to load
dynamically caml bytecode.
-- 
Rémi Vanicat
vanicat@labri.u-bordeaux.fr
http://dept-info.labri.u-bordeaux.fr/~vanicat
-------------------
Bug reports: http://caml.inria.fr/bin/caml-bugs  FAQ: http://caml.inria.fr/FAQ/
To unsubscribe, mail caml-list-request@inria.fr  Archives: http://caml.inria.fr